Hey,
I've noticed a strange bug in 2.4.13-ac2, 3, and 5. (It wasn't
there in 2.4.12-ac2) Namely, shortly after starting X, the reported
amount of cached memory spikes to somewhere around 18 hexabytes.
Needless to say, I don't have this much memory. :-)
The problem first occured on my pre-hardware-upgraded system, and
continues to occur on the post-upgrade system. The first system was:
ASUS A7V
900MHz Athlon T-bird
768MB RAM
linux-2.4.13-ac2/ac3 (UP, CONFIG_HIGHMEM was off)
The post-upgrade system is:
Tyan Tiger MP S2460
2x 1.4GHz Athlon MP 1600+
1024MB RAM
linux-2.4.13-ac5 (SMP, CONFIG_HIGHMEM was on)
/proc/meminfo on the SMP system is:
total: used: free: shared: buffers: cached:
Mem: 1053028352 812806144 240222208 8007680 385331200 18446744073599795200
Swap: 263168000 0 263168000
MemTotal: 1028348 kB
MemFree: 234592 kB
MemShared: 7820 kB
Buffers: 376300 kB
Cached: 4294860112 kB
SwapCached: 0 kB
Active: 453608 kB
Inact_dirty: 191808 kB
Inact_clean: 0 kB
Inact_target: 209700 kB
HighTotal: 131008 kB
HighFree: 2040 kB
LowTotal: 897340 kB
LowFree: 232556 kB
SwapTotal: 257000 kB
SwapFree: 257000 kB
It doesn't seem to cause any problems, as the system continues to
function normally. (Though I've not really tested to see if there's any
performance hit associated with the phenomenon) Either way, I'm guessing
it's not what's supposed to be happening. (As much as I'd love to have
that much cache)
If you'd like any further information, please let me know. Oh,
and please CC me as I'm not on this list.
Thanks,
Chris
--------------------------------
Chris Tracy
System/Network Administrator
Engineering Design Center
Santa Clara University
"Wherever you go, there you are."
On Thu, Nov 01, 2001 at 12:23:31AM -0800, Chris Tracy wrote:
> Hey,
>
> I've noticed a strange bug in 2.4.13-ac2, 3, and 5. (It wasn't
> there in 2.4.12-ac2) Namely, shortly after starting X, the reported
> amount of cached memory spikes to somewhere around 18 hexabytes.
> Needless to say, I don't have this much memory. :-)
Rik posted a patch yesterday to fix this.
mvg,
Alex
--
Eidetica [email protected]
Kruislaan 400 tel +31 20 888 4090 fax 4001
NL 1098 SM Amsterdam Home machine: myp.ath.cx
http://www.eidetica.com/ Alexander Kellett (Lypanov)
vim: tw=70 cindent!
On approximately Thu, Nov 01, 2001 at 09:26:39AM +0100, Alexander Kellett wrote:
> > Hey,
> >
> > I've noticed a strange bug in 2.4.13-ac2, 3, and 5. (It wasn't
> > there in 2.4.12-ac2) Namely, shortly after starting X, the reported
> > amount of cached memory spikes to somewhere around 18 hexabytes.
> > Needless to say, I don't have this much memory. :-)
>
> Rik posted a patch yesterday to fix this.
>
Here is the mentioned patch again.
--- linux-2.4.13-ac5/fs/proc/proc_misc.c.blkpg Wed Oct 31 13:09:51 2001
+++ linux-2.4.13-ac5/fs/proc/proc_misc.c Wed Oct 31 13:12:27 2001
@@ -140,7 +140,9 @@
{
struct sysinfo i;
int len;
- int pg_size;
+ unsigned int cached;
+
+ cached = atomic_read(&page_cache_size) - atomic_read(&shmem_nrpages);
/*
* display in kilobytes.
@@ -149,14 +151,12 @@
#define B(x) ((unsigned long long)(x) << PAGE_SHIFT)
si_meminfo(&i);
si_swapinfo(&i);
- pg_size = atomic_read(&page_cache_size) - i.bufferram ;
-
len = sprintf(page, " total: used: free: shared: buffers: cached:\n"
"Mem: %8Lu %8Lu %8Lu %8Lu %8Lu %8Lu\n"
"Swap: %8Lu %8Lu %8Lu\n",
B(i.totalram), B(i.totalram-i.freeram), B(i.freeram),
B(i.sharedram), B(i.bufferram),
- B(pg_size), B(i.totalswap),
+ B(cached), B(i.totalswap),
B(i.totalswap-i.freeswap), B(i.freeswap));
/*
* Tagged format, for easy grepping and expansion.
@@ -184,7 +184,7 @@
K(i.freeram),
K(i.sharedram),
K(i.bufferram),
- K(pg_size - swapper_space.nrpages),
+ K(cached - swapper_space.nrpages),
K(swapper_space.nrpages),
K(nr_active_pages),
K(nr_inactive_dirty_pages),
--
Linux, the choice | <doogie> Thinking is dangerous. It leads
of a GNU generation -o) | to ideas. -- Seen on #Debian
Kernel 2.4.14-pre6 /\ |
on a i586 _\_v |
|